Prey encounters and spatial memory influence use of foraging patches in a marine central place forager

R codes to reproduce the analysis presented in Iorio-Merlo et al. 2021 The scripts provided allow to repeat the analysis presented in the publication starting from the data available on dryad.

R and packages

The code was developed using R version 4.1.1
The pakages used are: adehabitatHR, adehabitatLT, flextable, geosphere, ggplot2, ggspatial, ggsn, lubridate, mefa, momentuHMM, pracma, raster, REdaS, rgdal, roll, sp, sf, tidyverse.

Scripts

In the folder R there are 11 scripts which should be run in the order they are numbered. Scripts are used to prepare the data, run the analysis and generate outputs.

• 01. Model 1 - Data preparation and dive batches
  This code is used for the initial filtering of the dive data and the grouping of dives in 5 dive batches. It filters dives that are within 2km of the haul-out sites. Then it filters for foraging trips (i.e. round-trips to the same haul-out site longer than 12 hours) and removes trips that occurred in the first week post-tagging.

• 02. Model 1 - Dive batches state classification
  This script is used to run the first HMM model using momentuHMM package. The analysis is run using these steps: HMM data preparation (step length and turning angle), selection of model initial parameters, model running.

• 03. Repeatability - Overlap between consecutive months
  This script uses the kernelUD() function from the adehabitatHR package to calculate the utilizations distribution (UD) of harbour seals while they are in an ARS state (output of model 1) for two consecutive months. As a first step we use the methods and codes from Lascelles et al. (2016) to estimate the most appropriate smoothing parameters h. We then estimate the overlap (Bhattacharaya affinity) of the distribution between the two months using the overlaphr(). we also obtain a null distribution of affinity index comparing the overlap between different individuals.

• 04. Spatial memory - Creating memory raster (a & b) This script uses the output from model 1 to create raster maps of areas where the seals have displayed ARS behaviour. It uses the mid-point of each dive batch to assign dive batches to a 1x1 km grid. For each grid cell it then calculates the proportion of dive batches classifed as ARS. Cells with no dive batches are assigned a mean value. The two codes are to create (a) the coavariates for model 2 and (b) the covariate for model 3.

• 05. Model 2 - Influence of spatial memory This script follows the same steps of script 2 to run the HMM model. Additional steps are the import of the raster maps generated in script 4a and a covariates selection step, where the inclusion of the covariate is assessed.

• 06. Prey encounters - At-sea accelerometer data extraction
  This short script is used to split the large accelerometer data into smaller datasets, one for each seal trip.

• 07. Prey encounters - PrCA axis standard deviation threshold
  This script is used to calculate an acceleration threshold to be used in the identification of Prey catch attempts in script 07. Prey encounters - Processing accelerometer data. The code takes the standard deviation of the dynamic acceleration over a moving window of 1.5 seconds and uses a kmean cluster analysis to define a "high" and "low" threshold. See Viviant et al. 2009, Cox et al. 2018

• 08. Prey encounters - Accelerometer processing functions
  This scripts reports a series of functions used to extract swimming effort, prey catch attempts (2 methodologies, Viviant et al.2010 and Brasseur et al. 2012) from the accelerometer.

• 09. Prey encounters - Processing accelerometer data
  This scripts uses the functions described in script 6 and other functions to calculate dive and accelerometers parameters for each dive in the dataset.

• 10. Model 3 - Data preparation and dive batches
  This script is very similar to script 1. It filters the data and groups the dives in 5-dive batches. For each batch the accelerometer and dive parameters are a mean of the values of the 5 dives.

• 11. Model 3 - Dive batches state classification and covariate influence
  This script follows the same steps of script 2 to run the HMM model. Additional steps are the import of the raster maps generated in script 4b and a covariates selection step, where the inclusion of either or both covariates is assessed.

License

GNU GENERAL PUBLIC LICENSE, Version 3, 29 June 2007 (GNU GPLv3)

References

Cox, S. L., et al. (2018). "Processing of acceleration and dive data on‐board satellite relay tags to investigate diving and foraging behaviour in free‐ranging marine predators." Methods in Ecology and Evolution 9(1): 64-77. doi.org/10.1111/2041-210X.12845

Brasseur, S., et al. (2012). "Habitat Preferences of Harbour Seals in the Dutch Coastal Area: Analysis and Estimate of Effects of Offshore Wind Farms (Report No. OWEZ R 252 T1 20120130 C043-10)", IMARES - Wageningen UR, Noordzeewind: 58.

Lascelles, B.G., et al. (2016). "Applying global criteria to tracking data to define important areas for marine conservation." Diversity and Distribution 22(4): 422-431. doi.org/10.1111/ddi.12411

McClintock, B.T. and Michelot, T. (2018). "momentuHMM : R package for generalized hidden Markov models of animal movement." Methods in Ecology and Evolution 9(6): 1518-1530. doi.org/10.1111/2041-210X.12995

Viviant, M., et al. (2010). "Prey capture attempts can be detected in Steller sea lions and other marine predators using accelerometers." Polar Biology 33(5): 713-719. doi.org/10.1007/s00300-009-0750-y